Re: Status and plans for glibc 2.37 release on February 1st 2023

[Adhemerval Zanella Netto <adhemerval.zanella@linaro.org>]

On 31/12/22 15:11, Carlos O'Donell via Libc-alpha wrote:
> Community,
> 
> I am volunteering as the release manager for glibc 2.37.
> 
> Starting with glibc 2.38 I would like to offer my time to mentor
> someone who would like to be a release manager for 2.38 (releasing
> August 1st 2023). I will make this offer again after the 2.37
> release.
> 
> Looking at the calendar the plan looks like this:
> 
> - Monday patch review switches to reviewing critical bug list and
>   ABI issues.
>  - Restarts January 9th.
> 
> - January 2-6 is slushy ABI freeze.
>  - No new ABI additions.
>  - ABI corrections should be reviewed, RM approval not required.
>   - Status of _r_debug COPY relocations is something we should look at.
>     https://patchwork.sourceware.org/project/glibc/list/?series=13978
>  - Liberal bug fixes.
>   - Status of pthread cond missed wakup is critical for this release.
>     - I will be making specific time to review the current status.
>       https://patchwork.sourceware.org/project/glibc/list/?series=13224
>       https://patchwork.sourceware.org/project/glibc/list/?series=12876
> 
> - January 9-31st is hard ABI freeze.
>  - Bug fixing.
>  - No ABI changes without RM discussion.
>  - Higher risk bug fixes should be discussed with community.
> 
> - February 1st
>  - glibc 2.37 release.
> 
> Please raise any concerns you have with this plan.
> 

I have added three fixes I would like for 2.37:

  * Fix remaining -Os/-O1 compile issues [BZ #19444] [1]: this is a long
    standing that triggers compiler warnings not enabled on all optimization
    levels.  It would be good if we can build glibc with multiple optimization
    options to check for such issues.

  * elf: Fix GL(dl_phdr) and GL(dl_phnum) for static builds [BZ #29864] [2]: 
    this issue is preventing the bootstrap on some archs, like alpha.

  * Remove --with-default-link option [3]: this prevents the vtable hardening
    when using --with-default-link=yes (and it also simplifies a bit the
    build setup).

I also added the 'Optimize posix_spawn signal setup with clone3' as desirable
feature.  Both the aarch64 and powerpc64 its are done and it leverages newer
kernel features to speed up posix_spawn.

[1] https://patchwork.sourceware.org/project/glibc/list/?series=15572
[2] https://patchwork.sourceware.org/project/glibc/list/?series=15662
[3] https://patchwork.sourceware.org/project/glibc/list/?series=15551
[4] https://patchwork.sourceware.org/project/glibc/list/?series=13901